基于K-Means聚类模糊算法的学生特征聚类研究

Research on Student Feature Clustering Based on K-Means Clustering Fuzzy Algorithm

摘要: 聚类分析是一种关于大量数据挖掘的重要技术。而对于学生存在的个性化差异，各自会表现出不同的特征。以不同学生所体现的特征来实现在教学过程中对学生因材施教，其特征类型的统计是一种大量数据处理的操作。所以，对于这种维度高、数据集大的海量数据，可以利用K-means算法结合数学模型给出聚类的学生特征，再引入模糊数学中的隶属度，来提取出更具解释性的聚类特征。首先，介绍了K-means算法的思想原理并分析其优缺点，并引入K-means++算法；其次，针对初始聚类中心点的选取和K值的确定；然后通过对大学生计算机专业的课程成绩为例，进行聚类分析，其实验结果表明，基于K-means聚类模糊算法的学生特征聚类相比基本的K-means或系统聚类，聚类结果上体现出更好的科学性和解释性；最后，对K-means算法技术关于学生学习成绩和效率进行展望。同时，本文利用蒙特卡罗和用频数来确定隶属度的想法，是团队首创，确保了创新性。

Abstract: Cluster analysis is an important technique for large-scale data mining. And for the personalized dif-ferences that students have, they will exhibit different characteristics. Teaching students according to their aptitude in the teaching process based on the characteristics reflected by different students is a large-scale data processing operation. Therefore, for such high-dimensional and large datasets, K-means algorithm can be combined with mathematical models to provide clustered student features, and membership degree in fuzzy mathematics can be introduced to extract more explanatory clustering features. Firstly, the concept and principle of the K-means algorithm were introduced, and its advantages and disadvantages were analyzed, followed by the introduction of the K-means++ algorithm; secondly, regarding the selection of initial clustering center points and the determination of K values; then, taking the course grades of college students majoring in computer science as an example, clustering analysis was conducted. The experimental results showed that the student feature clustering based on K-means clustering fuzzy algorithm showed better scientific and interpretive results compared to basic K-means or system clustering; Finally, the prospects of K-means algorithm technology for students’ academic performance and efficiency are presented. Also, the idea of using Monte Carlo algorithm and using frequency to determine affiliation in this paper is a first for the team and ensures innovation.